Rapid Identification of Legionella Spp. by MALDI-TOF MS Based Protein

G Model SYAPM-25346; No. of Pages 5 ARTICLE IN PRESS

Systematic and Applied Microbiology xxx (2010) xxx–xxx

Contents lists available at ScienceDirect

Systematic and Applied Microbiology

journal homepage: www.elsevier.de/syapm

Rapid identiﬁcation of Legionella spp. by MALDI-TOF MS based protein mass ﬁngerprinting

Valeria Gaia a,b,∗, Simona Casati a,b, Mauro Tonolla a,c a Cantonal Institute of Microbiology, Bellinzona, Switzerland b National Reference Centre for Legionella, Bellinzona, Switzerland c Microbial Ecology, Microbiology Unit, Plant Biology Department, University of Geneva, Switzerland article info abstract

Keywords: A set of reference strains representing 38 different Legionella species were submitted to Whole Cell Mass Legionella spp. Spectrometry (WCMS) with MALDI-TOF. MALDI-TOF MS The dendrogram computed from strain mass spectral patterns obtained by WCMS was compared to the mip phylogenetic tree obtained from macrophage infectivity potentiator (mip) sequences. The trees inferred Identification from these two methods revealed significant homologies. Sequencing Taxonomy Using 453 Legionella isolates previously characterized by genotyping, it was possible to create species- specific SuperSpectra, using appropriate sets of spectral masses, allowing unambiguous differentiation and identification of the most frequently isolated Legionella species. These SuperSpectra were tested for their suitability to identify Legionella strains isolated from water samples, cooling towers, potting soils and patient specimens deposited at the Swiss National Reference Centre for Legionella and previously identified by molecular methods such as mip gene sequencing. 99.1% of the tested strains isolated from the environment could be correctly identified by comparison with the new SuperSpectra. The identification of Legionella spp. by MALDI-TOF MS is rapid, easy to perform and has the advantage of being time- and cost-saving, in comparison to sequence-based identification. © 2010 Elsevier GmbH. All rights reserved.

Introduction infectivity potentiator), rpoB (␤-subunit of DNA-dependent RNA polymerase) or gyrA (gyrase A) [7,11,22], as well as immunolog- Species of Legionella cause severe forms of pneumonia called ical tests with species-specific antibodies (serotyping) [13–16]. Legionnaires’ disease. The first outbreak occurred in 1976 in a hotel L. pneumophila and L. anisa can be rapidly identified using com- in Philadelphia, where the state convention of the American Legion mercially available agglutination tests that show a very high was taking place: of 182 legionnaires who contracted the disease, specificity (100% for L. pneumophila and 99.5% for L. anisa) 29 died. Legionella pneumophila was reported as the causal agent [17,29]. For all other species, identification is usually per- for the first time a few months later [10,26]. Legionella spp. are formed by mip sequencing [28]. An internet-available database naturally present in water and soil, in association with amoebae (www.ewgli.org) allows fast retrieval of known, deposited DNA and other protozoa and in biofilms [8,30]. The family Legionel- sequences; sequencing, however, not only requires a considerable laceae includes 53 recognized species in one genus, 20 of which amount of time and work but is also coupled with relatively high are considered to be human pathogens [2,9]. The majority of cases costs. of Legionnaires’ diseases in Europe are caused by L. pneumophila It is almost impossible to distinguish different Legionella species serogroup 1 [1]. based on the appearance of the colonies on agar media. A rapid The reference methods used for the identification of Legionella and inexpensive technique for direct screening of colonies would spp. include sequencing of specific genes, such as mip (macrophage be very useful, in particular when studying environmental samples from cooling towers, potting soils, and composts, for example, where many non-L. pneumophila species, such as L. bozemanii, L. longbeachae, L. cincinnatiensis, L. jamestowniensis, L. micdadei and L. Abbreviations: MALDI-TOF MS, Matrix Assisted Laser Desorption Ionization – oakridgensis, can be simultaneously present [4,5,20,23,32]. Time Of Flight Mass Spectrometry; SARAMISTM, Spectral Archive And Microbial Whole Cell MALDI-TOF MS (WCMS) is increasingly used for the Identification System. ∗ identification of bacterial strains [3,6,18,19]. Only small amounts Corresponding author at: Cantonal Institute of Microbiology, Bellinzona, of cells are required; the procedure can be performed using the Switzerland. Tel.: +41 91 814 60 18; fax: +41 91 814 60 19. E-mail address: [email protected] (V. Gaia). biomass from single colonies. A recent paper showed the usefulness

Please cite this article in press as: V. Gaia, et al., Rapid identiﬁcation of Legionella spp. by MALDI-TOF MS based protein mass ﬁngerprinting, Syst. Appl. Microbiol. (2011), doi:10.1016/j.syapm.2010.11.007 G Model SYAPM-25346; No. of Pages 5 ARTICLE IN PRESS 2 V. Gaia et al. / Systematic and Applied Microbiology xxx (2010) xxx–xxx

Table 1 isolated during the last 10 years by the National Reference Centre Reference strains obtained from culture collections and used in this study, and (NRC) for Legionella in Bellinzona. accession numbers for mip sequences. To avoid variations in protein spectra that could result from the Species Strain No. NCBI sequence different ecological provenances of the isolates, all strains were accession number purified and cultivated on buffered charcoal yeast extract agar (mip) (BCYE) (bioMérieux, Geneva, Switzerland) at 36 ◦C for 48–72 h. 1 Legionella adelaidensis ATCC 49625 U91606 2 Legionella anisa ATCC 35292 U91607 Identification of the Legionella spp. studied 3 Legionella beliardensis ATCC 700512 AF047756 4 Legionella birminghamensis ATCC 43702 U91608 5 Legionella bozemanii ATCC 33217 U91609 Sequences of the mip gene of the 38 reference strains were 6 Legionella brunensis ATCC 43878 U92227 obtained from NCBI GenBank (Table 1). Isolates of L. pneumophila 7 Legionella cherrii ATCC 35252 U91635 and L. anisa (243 strains) were identified by the agglutination test 8 Legionella cincinnatiensis ATCC 43753 U91636 9 Legionella drozanskii ATCC 700990 AF148983 (Slidex Legionella, bioMérieux, Switzerland). Isolates for which 10 Legionella dumoffii ATCC 33279 U91637 no agglutination was observed (210 strains) were analyzed by 11 Legionella erythra ATCC 35303 U92203 mip-sequencing as previously described [28] and the sequences 12 Legionella fallonii ATCC 700992 AF148987 compared to those included in the mip identification database 13 Legionella feelei ATCC 35072 U92205 (www.ewgli.org). 14 Legionella geestiana ATCC 49504 FJ534536a 15 Legionella gormanii ATCC 33297 U91638 16 Legionella gresilensis ATCC 700509 AF047755 Sequence data analysis of reference strains 17 Legionella hackeliae ATCC 35250 U92207 18 Legionella impletisoli DSM 18493 AB233217 The nucleotide (nt) sequences of the mip gene of the reference 19 Legionella israelensis ATCC 43119 U92208 20 Legionella jamestowniensis ATCC 35298 U92228 strains were downloaded from NCBI (Table 1). For most species 21 Legionella jordanis ATCC 33623 U92209 sequences of 557 nt were obtained, whereas for L. yabuuchiae and 22 Legionella lansingensis ATCC 49751 U92210 L. impletisoli, only 417 nt were available. L. geestiana, which failed to 23 Legionella londiniensis ATCC 49505 U92229 produce an amplicon of the correct size (only 194 nt were available) 24 Legionella longbeachae ATCC 33462 X83036 [28] was not included in the alignment. 25 Legionella maceachernii ATCC 35300 U92211 26 Legionella micdadei ATCC 33218 FJ534537 The sequence similarities of strains were analyzed and phyloge- 27 Legionella moravica ATCC 43877 U92212 netic trees were constructed, using the Neighbor joining algorithm, 28 Legionella oakridgensis ATCC 33761 U92214 included in the MEGA 4 package. Confidence levels of inferred rela- 29 Legionella parisiensis ATCC 35299 U92215 tionships were estimated following 1000 bootstrap iterations [24]. 30 Legionella pneumophila ATCC 33152 AJ878849 31 Legionella quinlivanii ATCC 43830 U92217 32 Legionella rubrilucens ATCC 35304 U92218 MALDI-TOF MS 33 Legionella sainthelensi ATCC 35248 U92219 34 Legionella steigerwaltii ATCC 35302 U92223 A small amount (approx. 0.5 ␮l) of material was taken from 35 Legionella taurinensis ATCC 700508 AF022342 freshly grown colonies and transferred with a plastic loop into 36 Legionella tucsonensis ATCC 49180 U92224 TM 37 Legionella wadsworthii ATCC 33877 U92225 a well of a 48-well stainless steel FLEXImass target plate 38 Legionella yabuuchiae DSM 18492 AB233215 (Shimadzu Biotech, Kyoto, Japan). Analyses were run in dupli-

a Partial sequence of 194 nt. cate by spotting a colony into two different wells. The bacteria were overlaid with 0.5 ␮l of matrix solution containing 75 mg/ml 2,5dihydroxybenzoic acid (DHB) in acetonitrile/ethanol/water of WCMS for the rapid and inexpensive identification of twenty-one (1:1:1) supplemented with 3% trifluoroacetic acid and allowed to human pathogenic Legionella species [27]. dry at room temperature until the DHB crystals became visible. The SARAMIS database (Spectral ARchiving And Microbial Iden- MS analyses were performed in positive linear mode in the tification System, AnagnosTec GmbH, Potsdam, Germany) was range of 2000–20,000 masses-to-charge ratio (m/z), with an AXIMA developed for the rapid identification of bacterial strains [21]. ConfidenceTM Mass Spectrometer (Shimadzu Biotech). The database provides identification spectra for the majority of Average profile spectra, fulfilling the quality criteria (minimum clinically important, pathogenic species, yet Legionella spp. are intensity of the base peak: 10 mV; minimum signal to noise ratio: presently not enough represented. 10; minimum acceptable base peak resolution: 300), were collected The aim of this study is to compare WCMS results obtained by from 20 laser shot cycles. For each sample, 2 × 50 averaged profile SARAMIS with those obtained by mip gene phylogenetic analysis spectra were stored as mass spectra and used for analysis. and to confirm the robustness of the clustering obtained by WCMS. Furthermore, we wanted to complement the WCMS database with Analysis of the spectra the spectra of additional Legionella species. The peak lists of each strain were exported to the SARAMISTM Materials and methods software package (Anagnostec, Potsdam, Germany) and submitted to single-linkage cluster analysis, using the Dice coefficient, to pro- Bacterial strains duce dendrograms of strain similarities (0.08% error, range from 2000 to 20,000 m/z). Type strains from reference culture collections representing 38 The SARAMISTM software uses SuperSpectra for sample iden- different Legionella species (Table 1) were used in this study. tification. SuperSpectra are computed from the mass spectra of We also included in our analysis 453 isolates from clinical spec- multiple isolates belonging to the same species, by identifying imens and environmental samples, such as water, cooling towers consensus mass signals recorded in high frequency in the set and soils that were split into two sets: 216 isolates were used to of mass spectra, followed by weighting of the consensus peaks create species-specific SuperSpectra, whereas the remaining 237 in accordance with their specificities. By this procedure, reliable served to test the performance of the new SuperSpectra (Table 2). species-specific mass patterns are extracted that can be used as These strains belong to the 14 Legionella species most frequently reference data for automated ICMS-identification [21].

Please cite this article in press as: V. Gaia, et al., Rapid identiﬁcation of Legionella spp. by MALDI-TOF MS based protein mass ﬁngerprinting, Syst. Appl. Microbiol. (2011), doi:10.1016/j.syapm.2010.11.007 G Model SYAPM-25346; No. of Pages 5 ARTICLE IN PRESS V. Gaia et al. / Systematic and Applied Microbiology xxx (2010) xxx–xxx 3

Table 2 Origin of the isolates used for the creation and validation of the Legionella SuperSpectra (SSP).

Speciesa Type strains Patients Water Cooling towers Soil Total strains Strains used for Strains used for Discrepancyb creation of SSP validation of SSP

L. pneumophila 1 60 60 50 55 225 75 150 L. bozemanii 1 2 14 50 67 57 10 L. londiniensis 1 11 6 11 29 14 15 L. taurinensis 124251312 L. anisa 1 13 4 18 10 8 L. sainthelensi 1411016133 L. micdadei 1 2 1 10 14 4 10 1 L. longbeachae 15 51165 L. oakridgensis 1 2 5 9 17 6 11 L. cincinnatiensis 12 473 4 1 L. dumofﬁi 132633 L. birminghamensis 1113633 L. jamestowniensis 13 2642 L. rubrilucens 15 651

Total 14 66 146 69 159 453

a Identifications determined by conventional techniques. b One isolate identified by mip as L. micdadei was identified as L. oakridgensis by WCMS and one identified by mip as L. cincinnatiensis was identified as L. bozemanii by WCMS.

SuperSpectra for each of the 14 most frequently isolated species SuperSpectra creation and validation of Legionella were created after identification of species-specific masses and integrated in the SARAMISTM database for future rapid The MALDI-TOF MS spectral data resulting from the analysis of identification of new isolates. SuperSpectra were created only for 216 isolates were processed using the SARAMISTM database soft- species, for which at least three isolates were available, according ware to identify species-specific masses that were then used to to the Manufacturer’s instructions (Table 2). generate SuperSpectra for the identification of 14 Legionella species (Table 2). The robustness of the new SuperSpectra was tested by sub- jecting 237 isolates to the standardized automated identification Results procedure used for other clinical isolates. All but two environmental isolates could be identified using the new SuperSpectra with All strains could be analyzed by WCMS and useful spectra could 99.9% confidence. The two isolates for which discrepant results be produced. For the L. geestiana reference strains, only a 194 were obtained corresponded to two isolates identified by mip as L. nt sequence was available in the NCBI database, because for this micdadei and L. cincinnatiensis; SARAMISTM identified those isolates species mip cannot be amplified using standard primers [28]; thus, as L. oakridgensis and L. bozemanii, respectively. it was not included in the phylogenetic tree.

Discussion Comparison of mip and WCMS analysis of reference strains from culture collection For all reference strains of the 38 tested species of Legionella, spectra were produced, using standard extraction and analysis con- In the dendrogram based on the mass spectral patterns (Fig. 1A), ditions. isolates belonging to the same species clustered consistently in Seven groups were previously described by other authors using the same groups. The threshold level of the inter-species varia- phylogenetic analysis of genes, such as rnpB [31], rpoB [22] and tion was approximately 60% similarity, with the exception of two 16S rRNA [12,25]. Correspondingly, the WCMS analysis reflects clusters: L. taurinensis, L. erythra and L. rubrilucens (referred as the the phylogenetic classification inferred by mip sequence analy- red-fluorescent group); and L. anisa, L. bozemanii, L. parisiensis, L. sis. The blue-fluorescent group observed with WCMS was identical tucsonensis, L. gormanii, L. cherrii, L. wadsworthii, L. steigerwalti and to that seen in phylogenetic trees constructed with mip and rnpB L. dumoffii (referred as the blue-fluorescent group). Nevertheless, [28,31], whereas the same group was split into two separate clades, the variation of the spectra for species within these two groups using rpoB and 16S rRNA sequence analysis [12,22,25]. The red- allowed good differentiation at the species level (mass fingerprint fluorescent cluster and the group L. micdadei/L. maceachernii were similarity thresholds of 40% and 50% for the for the blue and the red already previously observed by other authors in phylogenetic trees fluorescent species, respectively). obtained from the rpoB, rnpB and 16S rRNA sequences [12,22,25]. In the mip tree, 7 clusters with bootstrap values higher than L. birminghamensis/L. quinlivanni cluster in one group in the phylo- 90% were observed: two comprising the blue- and red-fluorescent genetic trees constructed with the rnpB and 16S rRNA sequences groups, respectively; the cluster composed of L. sainthelensi, L. long- [12,25], but not in the one produced from rpoB sequence similarity beachae and L. cincinnatiensis; and also 4 small clusters composed analysis [22]. by L. beliardensis/L. gresilensis; L. birminghamensis/L. quinlivanii, L. In general, WCMS correlates well with the previously described micadei/L. maceachernii; and L. impletisoli/L. yabuuchiae (Fig. 1B). genotypic analyses. The only exception regards the cluster L. beliar- By comparing the WCMS dendrogram (Fig. 1A) with the phy- densis/L. quinlivanii and L. cincinnatiensis, which is separated from L. logenetic tree obtained from the mip sequences (Fig. 1B), one can longbeachae/L. sainthelensi by WCMS and not by sequence analysis. observe that 5 of the 7 clusters are supported by bootstrap values The high inter-species diversity observed within the WCSM greater than 90%. In the WCMS dendrogram, the group composed of spectra allowed reliable identification of all Legionella spp. L. beliardensis/L. gresilensis is split and L. cincinnatiensis is separated SuperSpectra were created only for the 14 most frequently iso- from L. longbeachae/L. sainthelensi. lated species, for which at least three isolates were available. All

Please cite this article in press as: V. Gaia, et al., Rapid identiﬁcation of Legionella spp. by MALDI-TOF MS based protein mass ﬁngerprinting, Syst. Appl. Microbiol. (2011), doi:10.1016/j.syapm.2010.11.007 G Model SYAPM-25346; No. of Pages 5 ARTICLE IN PRESS 4 V. Gaia et al. / Systematic and Applied Microbiology xxx (2010) xxx–xxx

Fig. 1. Dendrogram obtained by WCMS (A) compared to the phylogenetic tree (B) obtained by mip sequence analyses of 38 type strains from culture collections. tested strains were identified with very high confidence values by In conclusion, MALDI-TOF MS, coupled with the use of Super- the newly created SuperSpectra and only 2 discrepancies out of 237 Spectra constructed using the SARAMISTM database represents a isolates tested were observed. rapid means for the reliable identification of clinically and envi- Our study has two limitations. First, the different species ronmentally relevant species of Legionella. The technique allows were unevenly represented with regard to the number of strains. to perform a full analysis from a single colony in only few min- The most frequent species, however, represent the majority of utes, thus providing an inexpensive and rapid screening of a large Legionella strains isolated from clinical and environmental sam- number of colonies within a short time. ples in Switzerland by the NRC and for these SuperSpectra could be produced. Second, the application of cluster analysis based on Acknowledgment m/z values produces a somewhat arbitrary divisions of taxa which cannot be directly correlated with phylogenetic distances. The den- The authors are grateful to Guido Vogel (Mabritec, Basel), drogram obtained from the analysis of MALDI-TOF MS data, on the Martin Welker (AnagnosTec, Potsdam), and Orlando Petrini (ICM, other hand, correlates very well with the grouping produced by Bellinzona) for fruitful discussions and support. the genetic analysis of the mip sequences, indicating that species of Legionella can be reliably distinguished by protein fingerprint- References ing. Even though a high degree of intra-specific variation in mass [1] Bartram, J., Chartier, Y., Lee, J.V., Pond, K., Surman-Lee, S. (Eds.), Legionella patterns was observed for L. pneumophila, it was not possible to and the Prevention of Legionellosis, World Health Organization, Geneva, differentiate and separate strains according to their serogroups. Switzerland. [2] Brenner, D.J., Steigerwalt, A.G., McDade, J.E. (1979) Classification of the Legion- Further studies are needed to identify specific masses that could naires’ disease bacterium: Legionella pneumophila, genus novum, species nova, be useful for distinguishing serogroup 1 strains. On the other hand, of the family Legionellaceae, familia nova. Ann. Intern. Med. 90, 656–658. variation in mass spectra was observed within different strains [3] Bright, J.J., Claydon, M.A., Soufian, M., Gordon, D.B. (2002) Rapid typing of bacteria using matrix-assisted laser desorption ionisation time-of-flight mass belonging to a given species. This heterogeneity of the spectra spectrometry and pattern recognition software. J. Microbiol. Methods 48, of isolates belonging to the same species suggests that further 127–138. discrimination between serogroups or single strains should be pos- [4] Casati, S., Conza, L., Bruin, J., Gaia, V. (2009) Compost facilities as a reservoir of Legionella pneumophila and other Legionella species. Clin. Microbiol. Infect. 16, sible using WCMS. 945–947.

Please cite this article in press as: V. Gaia, et al., Rapid identiﬁcation of Legionella spp. by MALDI-TOF MS based protein mass ﬁngerprinting, Syst. Appl. Microbiol. (2011), doi:10.1016/j.syapm.2010.11.007 G Model SYAPM-25346; No. of Pages 5 ARTICLE IN PRESS V. Gaia et al. / Systematic and Applied Microbiology xxx (2010) xxx–xxx 5

[5] Casati, S., Gioria-Martinoni, A., Gaia, V. (2009) Commercial potting soils as [19] Hsieh, S.Y., Tseng, C.L., Lee, Y.S., Kuo, A.J., Sun, C.F., Lin, Y.H., Chen, J.K. (2008) an alternative infection source of Legionella pneumophila and other Legionella Highly efficient classification and identification of human pathogenic bacteria species in Switzerland. Clin. Microbiol. Infect. 15, 571–575. by MALDI-TOF MS. Mol Cell. Proteomics 7, 448–456. [6] Dugas, A.J., Murray, K.K. (2008) On-target digestion of collected bacteria for [20] Hughes, M.S., Steele, T.W. (1994) Occurrence and distribution of Legionella MALDI mass spectrometry. Anal. Chim. Acta 627, 154–161. species in composted plant materials. Appl. Environ. Microbiol. 60, 2003–2005. [7] Feddersen, A., Meyer, H.G., Matthes, P., Bhakdi, S., Husmann, M. (2000) GyrA [21] Kallow, W., Erhard, M., Shah, H.N., Raptakis, E., Welker, M. (2010) MALDI-TOF sequence-based typing of Legionella. Med. Microbiol. Immunol. 189, 7–11. MS for microbial identification: years of experimental development to an estab- [8] Fields, B.S. (1996) The molecular ecology of legionellae. Trends Microbiol. 4, lished protocol. In: Shah, H.N., Gharbia, S.E. (Eds.), Mass Spectrometry form 286–290. Microbial Proteomics, Wiley and Sons, Chichester, UK, pp. 255–277. [9] Fields, B.S., Benson, R.F., Besser, R.E. (2002) Legionella and Legionnaires’ disease: [22] Ko, K.S., Lee, H.K., Park, M.Y., Lee, K.H., Yun, Y.J., Woo, S.Y., Miyamoto, H., Kook, 25 years of investigation. Clin. Microbiol. Rev. 15, 506–526. Y.H. (2002) Application of RNA polymerase beta-subunit gene (rpoB) sequences [10] Fraser, D.W., Tsai, T.R., Orenstein, W., Parkin, W.E., Beecham, H.J., Sharrar, R.G., for the molecular differentiation of Legionella species. J. Clin. Microbiol. 40, Harris, J., Mallison, G.F., Martin, S.M., McDade, J.E., Shepard, C.C., Brachman, P.S. 2653–2658. (1977) Legionnaires’ disease: description of an epidemic of pneumonia. N. Engl. [23] Koide, M., Arakaki, N., Saito, A. (2001) Distribution of Legionella longbeachae and J. Med. 297, 1189–1197. other legionellae in Japanese potting soils. J. Infect. Chemother. 7, 224–227. [11] Fry, N.K., Harrison, T.G. (1998) An evaluation of intergenic rRNA gene sequence [24] Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M. (2001) MEGA2: molecular evolu- length polymorphism analysis for the identification of Legionella species. J. Med. tionary genetics analysis software. Bioinformatics 17, 1244–1245. Microbiol. 47, 667–678. [25] Lo Presti, F., Riffard, S., Meugnier, H., Reyrolle, M., Lasne, Y., Grimont, P.A., Gri- [12] Fry, N.K., Warwick, S., Saunders, N.A., Embley, T.M. (1991) The use of 16S ribo- mont, F., Benson, R.F., Brenner, D.J., Steigerwalt, A.G., Etienne, J., Freney, J. (2001) somal RNA analyses to investigate the phylogeny of the family Legionellaceae. Legionella gresilensis sp. nov. and Legionella beliardensis sp. nov., isolated from J. Gen. Microbiol. 137, 1215–1222. water in France. Int. J. Syst. Evol. Microbiol. 51, 1949–1957. [13] Helbig, J.H., Benson, R.F., Pelaz, C., Jacobs, E., Luck, P.C. (2007) Identification and [26] McDade, J.E., Shepard, C.C., Fraser, D.W., Tsai, T.R., Redus, M.A., Dowdle, W.R. serotyping of atypical Legionella pneumophila strains isolated from human and (1977) Legionnaires’ disease: isolation of a bacterium and demonstration of its environmental sources. J. Appl. Microbiol. 102, 100–105. role in other respiratory disease. N. Engl. J. Med. 297, 1197–1203. [14] Helbig, J.H., Bernander, S., Castellani Pastoris, M., Etienne, J., Gaia, V., Lauwers, S., [27] Moliner, C., Ginevra, C., Jarraud, S., Flaudrops, C., Bedotto, M., Couderc, C., Eti- Lindsay, D., Luck, P.C., Marques, T., Mentula, S., Peeters, M.F., Pelaz, C., Struelens, enne, J., Fournier, P.E. (2009) Rapid identification of Legionella species by mass M., Uldum, S.A., Wewalka, G., Harrison, T.G. (2002) Pan-European study on spectrometry. J. Med. Microbiol. 59, 273–284. culture-proven Legionnaires’ disease: distribution of Legionella pneumophila [28] Ratcliff, R.M., Lanser, J.A., Manning, P.A., Heuzenroeder, M.W. (1998) Sequence- serogroups and monoclonal subgroups. Eur. J. Clin. Microbiol. Infect. Dis. 21, based classification scheme for the genus Legionella targeting the mip gene. J. 710–716. Clin. Microbiol. 36, 1560–1567. [15] Helbig, J.H., Kurtz, J.B., Pastoris, M.C., Pelaz, C., Luck, P.C. (1997) Antigenic [29] Reyrolle, M., Ratat, C., Leportier, M., Jarraud, S., Freney, J., Etienne, J. (2004) Rapid lipopolysaccharide components of Legionella pneumophila recognized by mon- identification of Legionella pneumophila serogroups by latex agglutination. oclonal antibodies: possibilities and limitations for division of the species into Eur. J. Clin. Microbiol. Infect. Dis. 23, 864–866. serogroups. J. Clin. Microbiol. 35, 2841–2845. [30] Rowbotham, T.J. (1980) Preliminary report on the pathogenicity of Legionella [16] Helbig, J.H., Luck, P.C., Knirel, Y.A., Witzleb, W., Zahringer, U. (1995) Molecular pneumophila for freshwater and soil amoebae. J. Clin. Pathol. 33, 1179–1183. characterization of a virulence-associated epitope on the lipopolysaccharide of [31] Rubin, C.J., Thollesson, M., Kirsebom, L.A., Herrmann, B. (2005) Phylogenetic Legionella pneumophila serogroup 1. Epidemiol. Infect. 115, 71–78. relationships and species differentiation of 39 Legionella species by sequence [17] Helbig, J.H., Luck, P.C., Kunz, B., Bubert, A. (2006) Evaluation of the Duopath determination of the RNase P RNA gene rnpB. Int. J. Syst. Evol. Microbiol. 55, Legionella lateral flow assay for identification of Legionella pneumophila and 2039–2049. Legionella species culture isolates. Appl. Environ. Microbiol. 72, 4489–4491. [32] Steele, T.W., Moore, C.V., Sangster, N. (1990) Distribution of Legionella long- [18] Hettick, J.M., Kashon, M.L., Slaven, J.E., Ma, Y., Simpson, J.P., Siegel, P.D., Mazurek, beachae serogroup 1 and other legionellae in potting soils in Australia. Appl. G.N., Weissman, D.N. (2006) Discrimination of intact mycobacteria at the strain Environ. Microbiol. 56, 2984–2988. level: a combined MALDI-TOF MS and biostatistical analysis. Proteomics 6, 6416–6425.